Artificial intelligence for tailored further education and training to enhance employability in the face of increasing digitalisation in manufacturing within the context of Industry 4.0.

Background and Problem

Advancing digitalisation in manufacturing companies in the context of Industry 4.0 is changing processes and the demands placed on employees. Increasing automation is leading to rapid change in work processes and equipment, with competence requirements changing in different ways depending on the area of responsibility. This creates the need for new, individualised continuing training offerings. One example is collaborative robotics, a key technology in the digitalisation process with high training needs. The KIWI project aims to harness the potential of digitalisation to design a continuing training environment for collaborative robotics that can be experienced in virtual reality (VR) as well as used on PCs and smartphones. In addition, the project uses artificial intelligence (AI) to recommend tailored continuing education and training, specifically for vocationally qualified workers at shop-floor level in the manufacturing industry whose tasks will change significantly as a result of automation. To this end, a user interface is being developed that acts as a meta search engine for various online education platforms. It is intended to enable employees to find suitable continuing training offerings and to create an individual online portfolio. In this way, users gain easy access to relevant educational offerings and can drive their professional development in a self-determined manner.

Research question

• What effects does the use of virtual reality (VR) as a continuing training tool have on the effectiveness of learning and on technology acceptance among employees in collaborative robotics?

• How can AI-based algorithms be used to create individualised continuing training offerings in the manufacturing industry in order to meet the rapidly changing competence requirements brought about by digitalisation and automation?

Approach

The continuing training scenarios for collaborative robotics were developed, tested and adapted in an ongoing iterative process.

In parallel, the database for the AI was set up, comprising content from the European Qualifications Framework (EQF), curricula for initial, advanced and continuing training, and existing continuing education and training offerings. Building on this, the planning and implementation of the backend for the AI-supported continuing education and training recommendation took place in an agile development process.

Finally, both the continuing training scenarios and the continuing training advisory tool were to be evaluated with the target group of employees in the manufacturing industry.

Results

A virtual learning environment for collaborative robotics applications based on the Unreal Engine had already been implemented. It could be used on head-mounted displays, on PCs and on smartphones. Since it ran on a server with high graphics performance, only an internet browser was required for use on PCs and smartphones. A tutorial level introduced users to controlling the environment, handling collaborative robots and basic learning content about them. In a human–robot collaboration (HRC) scenario, the assembly of a small gearbox could be learnt as an example.

It was also expected that, by the end of the project, continuing training recommendations could be issued to users on the basis of an AI language model. For this purpose, a user interface was to be designed that invites interaction, is easy to use and creates a more personal level of engagement with users by means of a digital avatar.

Dissemination and Transfer Activities

On 21 November 2023, at the conference ‘KI² – Konkrete Ideen kooperativ & innovativ umsetzen’ (Implementing Concrete Ideas Cooperatively and Innovatively) in Nuremberg, the prototype of the digital avatar as the user interface for continuing training guidance based on the language model in the Unreal Engine (UE), together with the VR continuing training scenario, was presented to interested visitors at an exhibition stand and feedback was gathered.

Publications:

Konzept für eine Smartphone-basierte VR-Lernumgebung / Kenneth Rüstmann; Kirsten Tracht, In: wt Werkstattstechnik online. 9-2023, S. 341-346. https://doi.org/10.37544/1436-4980-2023-09-5